Risk of Pancreatic Adenocarcinoma in Chronic Pancreatitis D Malka, P Hammel, F Maire, P Rufat, I Madeira, F Pessione, P Lévy, P Ruszniewski

849

PANCREATIC DISEASE Gut: first published as 10.1136/gut.51.6.849 on 1 December 2002. Downloaded from Risk of pancreatic adenocarcinoma in chronic pancreatitis D Malka, P Hammel, F Maire, P Rufat, I Madeira, F Pessione, P Lévy, P Ruszniewski ......

Gut 2002;51:849–852

Background: The risk of pancreatic cancer in patients with chronic pancreatitis (CP) is difficult to assess. Previous studies, mostly case control studies or studies relying on data case registers, reported See end of article for authors’ affiliations relative risks varying from 2.3 to 18.5...... Methods: We studied a prospective, single centre, medical-surgical cohort of 373 consecutive patients (322 (86%) men, median age 40 years) with proven CP (alcoholic origin 85%) and a follow up of at Correspondence to: Professor P Ruszniewski, least two years (median follow up 9.2 years; range 2.0–34.8) in order to exclude pancreatitis reveal- Service de Gastro- ing pancreatic cancer. We calculated the age and sex standardised incidence ratio (SIR) as the ratio Entérologie, Hôpital of the number of observed cases of pancreatic cancer in this cohort to the number of expected cases, Beaujon, AP-Hp, 100 as provided by the French National Cancer Register. Boulevard du Général Results: Four cases of pancreatic adenocarcinoma (1.1% of patients) were observed in 3437 patient Leclerc, F-92118 Clichy Cedex, France; years (expected number of cases 0.15; SIR 26.7, 95% confidence interval (CI) 7.3–68.3; philippe.ruszniewski@ p=0.00002). In a second analysis in which patients lost to follow up were considered to be followed bjn.ap-hop-paris.fr up until the end point without having developed pancreatic adenocarcinoma (4762 patient years), SIR Accepted for publication was 19.0 (CI 5.2–48.8; p=0.00007). 25 June 2002 Conclusion: Patients with CP have a markedly increased risk of pancreatic cancer compared with the ...... general population.

he aetiology of pancreatic cancer remains largely elusive. x ray, computed tomography scan, ultrasonography, or Smoking, the only consistent environmental risk factor, is endoscopic ultrasonography; (2) moderate to marked associated with an approximately threefold increase in pancreatic ductal lesions on endoscopic retrograde or intra- T 1 15 the risk of pancreatic cancer, and less than 5% of cases of operative pancreatography (“Cambridge” criteria) ; (3) typi- pancreatic cancer are thought to be related to familial cal histology on an adequate surgical pancreatic specimen. CP http://gut.bmj.com/ (genetic) factors.2 was considered to be due to alcohol when alcohol intake Anecdotal case reports and short case control studies have exceeded 60 g/day for at least two years in the absence of other suggested that chronic pancreatitis (CP) is a risk factor for causes; CP was considered “idiopathic” when no cause was pancreatic cancer.3–8 Recently, several studies have reinforced found. The date of clinical onset of CP was defined as the date this hypothesis.9–14 However, the risk of pancreatic cancer in when the first manifestation clearly attributable to CP patients with CP varied widely from 2.311 to 18.514 in these occurred. The diagnosis of pancreatic adenocarcinoma had to studies, raising methodological concerns. In most, the diagno- be proved in all cases by histological examination of surgical on September 28, 2021 by guest. Protected copyright. sis of pancreatic cancer (mostly without histological confirma- specimens or fine needle biopsy. Special attention was paid to tion) or pancreatitis relied on data in inpatient, cancer, or death exclude pancreatic adenocarcinoma arising on intraductal registers,10 11 13 or even on a questionnaire12; both acute papillary mucin producing tumours. pancreatitis, unspecified pancreatitis, and CP were analysed in 10–13 9–13 Study design and data collection several series ; and most studies were retrospective, began 9 9 Patients were seen as outpatients at least once a year, or in 1946, or were multicentric, leading to heterogeneity and unscheduled (as outpatients or hospitalised) when sympto- possible biases. We thus aimed to determine the risk of pancre- matic or when their disease was complicated. Special attention atic adenocarcinoma in a prospective, single centre, medical- was paid to signs or symptoms suggestive of pancreatic cancer surgical cohort of consecutive patients with proven CP. (pain, weight loss, jaundice, abdominal mass). Abdominal imaging and laboratory tests were performed when necessary. PATIENTS AND METHODS Patients were considered lost to follow up when they failed to Patients attend our institution for more than one year. Person years All consecutive patients with suspected CP attending the (that is, the observation time contributed by one patient medical and surgical gastroenterological units of our institu- followed for one year) were calculated from the date of clini- tion between 1973 and the end point of this study (July 1997), cal onset of CP,and either the end point of the study, the last who fulfilled the diagnosis criteria of CP cited below during personal contact, the date of diagnosis of pancreatic adenocar- follow up, were prospectively studied. Patients with a follow cinoma, the date of total pancreatectomy (n=4), or death. up of less than two years or a diagnosis of pancreatic cancer Data were prospectively registered on a database. established during the first two years of follow up were excluded to rule out the possibility of pancreatitis revealing pancreatic adenocarcinoma.

Definitions ...... The diagnosis of CP was based on one or more of the follow- Abbreviations: CP, chronic pancreatitis; SIR, standardised incidence ing three criteria: (1) pancreatic calciﬁcations, as evidenced by ratio.

www.gutjnl.com 850 Malka, Hammel, Maire, et al

tion, 500 fulfilled the diagnostic criteria of CP and were Table 1 Clinical characteristics of the 373 patients prospectively studied. Among them, 127 patients (25%) with chronic pancreatitis (CP) followed up for less than two years after the onset of CP were Gut: first published as 10.1136/gut.51.6.849 on 1 December 2002. Downloaded from CP patients excluded from the present analysis. Table 1 presents the characteristics of the remaining 373 patients (3437 patient years). Median age at onset of CP (y) 40 (CI 17–67, range 5–86) Mean alcohol consumption of drinkers at the time of diagno- Median age at diagnosis of CP (y) 42 (CI 20–68, range 12–86) Median duration of follow up (y) 9.2 (CI 2.2–27.2, range sis of CP was 170 g/day. Mean cumulative tobacco consump- 2.0–34.8) tion of smokers at the time of diagnosis of CP was 35 pack Male (%) 322 (86) years. The causes of non-alcoholic CP were hereditary CP Alcoholic CP (%) 318 (85) (n=6), abdominal radiotherapy (n=3), inflammatory bowel Alcohol consumption (g/day)* 0 (%) 26 (7) disease (n=1), and Sjögren’s syndrome (n=1); CP was classi- <20 (%) 25 (7) fied as “idiopathic” in the remaining 40 (11%) cases. The 118 20–120 (%) 140 (40) patients (32%) lost to follow up did not differ from the >120 (%) 162 (46) remaining CP patients in their characteristics, except for a Tobacco consumption (pack years)†‡ longer duration of follow up (median follow up 10.0 years; CI 0 (%) 36 (12) <10 (%) 24 (8) 2.2–27.2; range 2.0–31.3). 10–40 (%) 159 (52) >40 (%) 88 (29) Risk of pancreatic adenocarcinoma Pancreatic calcifications (%) § 308 (83) Sixty one patients (16%) died within the follow up period: 10 Diabetes mellitus (%)§ 202 (54) patients died from liver disease, 11 from sepsis, 16 from mis- Insulin requirement (%) § 90 (24) cellaneous causes (hypoglycaemia, pulmonary embolism, Bouts of acute pancreatitis (%) § 207 (55) Pseudocysts (%) § 171 (46) tuberculosis, stroke), and seven from unknown causes. Duodenal stenosis (%) § 69 (18) Thirteen patients (3.5%) died from non-pancreatic cancer Splenic or portal vein occlusion (%) § 77 (21) (head and neck, six; liver, three; oesophagus, one; colon, one; Liver disease/cirrhosis (%) § 140 (38)/40 (11) stomach, one; bladder, one). Four patients (1.1%) died from Biliary liver disease (%) § 51 (14) pancreatic adenocarcinoma. None had a family history of Elective surgery for CP (%) § 222 (60) Uncomplicated CP (%)§¶ 42 (11) pancreatic cancer or hereditary CP (table 2). The expected Deaths (%) 61 (16) number of cases of pancreatic cancer was 0.15, yielding an SIR of 26.7 (CI 7.3–68.3; p=0.00002). The cumulative incidence of Available data for: *353 (95%) patients; †307 (82%) patients. pancreatic cancer was 1.1% at five years and 1.7% at 10 years. ‡Percentage totals exceed 100% because of rounding. §At any time during the course of CP. SIR was 19.0 (CI 5.2–48.8; p=0.00007) in a second analysis in ¶No acute pancreatitis, pseudocysts, duodenal stenosis, biliary liver which patients lost to follow up were considered to be disease, splenic or portal vein occlusion, need for pancreatic surgery, followed up until the end point without occurrence of pancre- or insulin. atic cancer (4762 patient years; expected number of cases of pancreatic cancer 0.21).

Statistical analysis DISCUSSION http://gut.bmj.com/ Differences were analysed using the Student’s t test or Mann- This single centre, medical-surgical, prospective cohort study Whitney U test as necessary for continuous data and the χ2 confirms that the risk of pancreatic cancer is markedly test or Fisher’s exact test as necessary for categorical data. We increased in CP patients compared with age and sex matched used age stratified (according to five year age groups) and sex controls. We found a much higher risk than in most case con- specific data on the incidence of cancer in France (French trol studies11 12 and population based cohort studies,10 13 several National Cancer Register, period 1983–87) to determine the of which having even challenged the hypothesis of an associ- expected number of cases of pancreatic cancer in the cohort. ation between CP and pancreatic cancer.16–20 The ratio of the observed number of cases of pancreatic cancer Assessing the true risk of pancreatic cancer in CP patients is on September 28, 2021 by guest. Protected copyright. in the cohort of CP patients to the expected number of cases hampered by several potential biases. Most case control stud- (standardised incidence ratio (SIR)) was used to estimate the ies reported risks that did not reach statistical significance relative risk. The 95% confidence interval (CI) for the SIR was because of wide CI values,16–20 or relied on interviews of calculated assuming that the observed cases of pancreatic patients with pancreatic cancer, thus suggesting the possi- cancer followed a Poisson distribution. Data were analysed bility of recall bias and subsequent overestimation of risk as with SAS 6.12 (SAS Institute Inc., Cary, North Carolina, USA). these patients may have been more sensitised than controls All statistical tests were two sided. Statistical significance was towards recalling past episodes of pancreatitis.12 set at p<0.05. Cohort studies stand and fall by the quality of the input data.21 22 The two large population based studies available RESULTS retrieved inpatients with either acute pancreatitis, recurrent Patient characteristics pancreatitis, CP, or even “unspecified” pancreatitis from Among the 567 patients with suspected CP attending the county10 or national13 registers from 1965 to 1983, raising con- medical and surgical gastroenterological units of our institu- cerns about the possibility of misclassification bias. The

Table 2 Characteristics of the four patients with chronic pancreatitis (CP) and pancreatic adenocarcinoma

Sex/ Alcohol intake Smoking Aetiology of CP Pancreatic Diabetes Symptoms revealing age (y) (g/day) (pack years) (duration (months)) calcifications mellitus Steatorrhea pancreatic cancer Outcome

F/43 20–120 10–40 Alcoholic (29) Yes No No Pain, weight loss, jaundice Died at 8 months F/54 >120 Not available Alcoholic (38) Yes Yes No Pain, weight loss, jaundice Died at 11 months M/77 <20 0 Idiopathic (55) No No Yes Pain, weight loss, jaundice, Died at 5 months duodenal stenosis M/38 20–120 <10 Alcoholic (103) Yes No Yes Pain, weight loss, jaundice Died at 14 months

www.gutjnl.com Pancreatic cancer in chronic pancreatitis 851 proportion of CP patients was unknown in the earlier study.10 and “the year of clinical onset was taken as the year in which In the latter,13 CP concerned men only 2.6 times more often the first unmistakable episode of pain occurred”.14 Overesti- than women, and was diagnosed at a mean age of 53.2 years, mation bias may also have resulted from recruitment of only Gut: first published as 10.1136/gut.51.6.849 on 1 December 2002. Downloaded from more than a decade older than the age at diagnosis in our inpatients with CP in all previous studies but one.14 study and in most other reports.923 Only 58% of a random The main limitation of our study was the high rate of sample (10%) of “CP” patients had definite diagnostic criteria patients lost to follow up. This is likely due to the long duration of CP (as ours), the remaining having either “recurrent pain” of follow up, the stringency of our definition of patients lost to (25%) or a “questionable” discharge diagnosis (17%).24 follow up, and the usual low compliance of alcoholic patients Misclassifying acute pancreatitis and CP may dilute the risk of which accounted for 85% of our patients. Patients lost to pancreatic cancer in CP patients and account for the intriguing follow up did not differ from the remaining CP patients in observed risk excess in patients with one single episode of their characteristics, except for a longer duration of follow up. 10 13 acute pancreatitis. The finding that the risk of pancreatic However, even though patients lost to follow up were consid- cancer at 10 years remained significant in the case of alcohol- ered to have been followed until the end point of the study ism but not in CP patients further suggests the possibility of without occurrence of pancreatic adenocarcinoma (4762 13 misclassification in subcohorts. patient years), the SIR of pancreatic cancer remained very The results of two previous cohort studies of CP patients are high (19.0). A similar approach was used in a previous study.9 914 14 in accordance with the present study. The latter wasinfact In contrast, hypothesising that (at least) one of the seven 9 an extension of a subcohort included in the former, to take unexplained deaths in this study were due in fact to into account the heterogeneity of CP patients recruited pancreatic cancer would result in an even higher SIR of pan- through seven centres in six countries, as attested by wide creatic cancer in CP patients. Another limitation of our study variations between the subcohorts for age at diagnosis of CP was the small number of observed cases of pancreatic cancer, (41.7 to 51.6 years), duration of follow up (5.9 to 9.4 years), precluding investigation of potential risk factors such as male sex (71% to 87%), alcoholic CP (68% to 82%), pancreatic smoking.1 calcifications (38% to 84%), diabetes (28% to 76%), pancreatic The mechanisms underlying the risk of pancreatic cancer in surgery (33% to 54%), deaths (14% to 47%), and pancreatic CP patients are unclear. Ductal epithelial hyperplasia, meta- cancer (1.3% to 2.7%). Recruitment of patients began in 1946 plasia and dysplasia, and Ki-ras gene mutations have been when it was virtually impossible to diagnose CP and to distin- described in CP patients, suggesting an oncogenetic multistep guish it from pancreatic cancer with certainty, and only 35% sequence.27 Chronic pancreatic inflammation may account for and 16% of patients underwent endoscopic retrograde the increased risk of pancreatic cancer in the course of CP pancreatography and computed tomography scan, 9 whether induced by environmental, genetic, or other causes, respectively. Moreover, the incidence of pancreatic cancer 28 29 dramatically increased several decades ago, suggesting the as is the case for other premalignant diseases. The duration possibility of detection bias in this historical study.25 of CP may influence the magnitude of the increase in the risk The finding that the risk of pancreatic cancer declined with of pancreatic cancer, as suggested by the estimated cumulative 10 13 11 risk of 40% to age 70 years found in patients with hereditary time in population based and case control studies raises 30 the possibility of a spurious association—pancreatic cancer CP. However, only 1.1% of CP patients in our study developed a pancreatic cancer, and CP may account for only 0.1%31 to masquerading as (or causing an obstructive form of) CP.This 12

5% of cases of pancreatic cancer compared with an attribut- http://gut.bmj.com/ misdiagnosis bias, which was more likely to occur in older 31 studies91013and in studies lacking stringent diagnostic criteria able risk of approximately 30% for smoking. For this reason, for CP,10–13 was minimised in our study and others91014 by a policy of pancreatic cancer screening in CP patients does not excluding CP patients in whom pancreatic cancer was seem to be advocated. In addition, diagnosing pancreatic can- diagnosed during the first two years of follow up. This policy cer in CP patients remains challenging, as pancreatic cancer led to exclusion of nearly half of the cases of pancreatic cancer was almost always diagnosed at an advanced stage despite 32 arising in CP patients in previous studies, further underlining close follow up in our study and others. the importance of this misdiagnosis bias.9 However, this two year exclusion policy—a fortiori if a one year cut off was ...... on September 28, 2021 by guest. Protected copyright. chosen11–13—might have been insufficient in studies in which Authors’ affiliations pancreatic cancer was not always proven by histological D Malka, P Hammel, F Maire, I Madeira, P Lévy, P Ruszniewski, examination as in our study9–11 13 to eliminate a significant Fédération Médico-Chirurgicale d’Hépato-Gastro-Entérologie, Hôpital Beaujon, Université Paris VII, Assistance Publique-Hôpitaux de Paris, proportion of cases corresponding to slower growing tumours, Clichy, France such as intraductal papillary mucin producing tumours. Of P Rufat, F Pessione, Cellule MSI, Hôpital Beaujon, Université Paris VII, note, in the most recent study, “obstructive” CP (that is, non- Assistance Publique-Hôpitaux de Paris, Clichy, France alcoholic, non-familial, non-idiopathic) accounted for 12% of cases, suggesting that a significant proportion of the REFERENCES population studied corresponded to obstructive CP caused by 1 Talamini G, Bassi C, Falconi M, et al. Alcohol and smoking as risk potentially premalignant lesions of the pancreas.14 factors in chronic pancreatitis and pancreatic cancer. Dig Dis Sci 1999;44:1303–11. Calculating person years from the date of diagnosis of 2 Chappuis PO, Ghadirian P, Foulkes WD. The role of genetic factors in CP9–11 13 rather than the date of onset of the disease as in our the etiology of pancreatic adenocarcinoma: an update. Cancer Invest study and others14 may cause overestimation bias by obviating 2001;19:65–75. 3 Lin JT, Wang TH, Chen DS, et al. Pancreatic carcinoma associated with the first years of the disease or may lead to illegitimate exclu- chronic calcifying pancreatitis in Taiwan: a case report and review of the sion of patients followed for less than two years since the date literature. Pancreas 1988;3:111–14. of diagnosis—particularly patients with a more indolent form 4 Farrow DC, Davis S. Risk of pancreatic cancer in relation to medical history and the use of tobacco, alcohol and coffee. Int J Cancer of CP.Our diagnostic criteria for CP were stringent but did not 1990;45:816–20. exclude painless CP, which accounts for 5–10% of cases of 5 La Vecchia C, Negri E, D’Avanzo B, et al. Medical history, diet and alcoholic CP and up to 50% of cases of non-alcoholic CP.23 26 pancreatic cancer. Oncology 1990;47:463–6. These patients were more likely to be recruited in previous 6 Haas O, Guillard G, Rat P, et al. Pancreatic carcinoma developing in chronic pancreatitis: a report of four cases. Hepatogastroenterology studies when pain occurred, not being due to CP but to 1990;37:350–1. pancreatic cancer, artificially increasing the risk of pancreatic 7 Misra SP, Thorat VK, Vij JC, et al. Development of carcinoma in chronic cancer within the first two years of follow up. Even in the most calcific pancreatitis. Int J Pancreatol 1990;6:307–12. 8 Kalapothaki V, Tzonou A, Hsieh CC, et al. Tobacco, ethanol, coffee, recent study in which the design closely resembled ours, the pancreatitis, diabetes mellitus, and cholelithiasis as risk factors for diagnosis was based on so-called “typical pancreatic” pain, pancreatic carcinoma. Cancer Causes Control 1993;4:375–82.

www.gutjnl.com 852 Malka, Hammel, Maire, et al

9 Lowenfels AB, Maisonneuve P, Cavallini G, et al. Pancreatitis and the 21 Lowenfels AB, Maisonneuve P. Pancreatic disease: does “itis” lead to risk of pancreatic cancer. International Pancreatitis Study Group. N Engl “oma”? Gastroenterology 1998;114:859–60. JMed1993;328:1433–7. 22 Andrén-Sandberg A. Unreliable pancreatitis epidemiology captures the 10 Ekbom A, McLaughlin JK, Karlsson BM, et al. Pancreatitis and wrong population. Gastroenterology 1998;114:860–1. Gut: first published as 10.1136/gut.51.6.849 on 1 December 2002. Downloaded from pancreatic cancer: a population-based study. J Natl Cancer Inst 23 Ammann RW, Akovbiantz A, Largiader F, et al. Course and outcome of 1994;86:625–7. chronic pancreatitis. Longitudinal study of a mixed medical-surgical series 11 Bansal P, Sonnenberg A. Pancreatitis is a risk factor for pancreatic of 245 patients. Gastroenterology 1984;86:820–8. cancer. Gastroenterology 1995;109:247–51. 24 Karlson BM, Ekbom A. Untitled. Gastroenterology 1998;114:861. 12 Fernandez E, La Vecchia C, Porta M, et al. Pancreatitis and the risk of 25 Gloeckler Ries LA, Hankey BF, Miller BA, et al, eds. Cancer statistics pancreatic cancer. Pancreas 1995;11:185–9. review 1973–88. Bethesda: Department of Health and Human Services, 13 Karlson BM, Ekbom A, Josefsson S, et al. The risk of pancreatic cancer 1991. (NIH publication No 91–2789). following pancreatitis: an association due to confounding? 26 Layer P, DiMagno EP. Early and late onset in idiopathic and alcoholic Gastroenterology 1997;113:587–92. chronic pancreatitis. Different clinical courses. Surg Clin North Am 14 Talamini G, Falconi M, Bassi C, et al. Incidence of cancer in the course 1999;79:847–60. of chronic pancreatitis. Am J Gastroenterol 1999;94:1253–60. 27 Apple SK, Hecht JR, Lewin DN, et al. Immunohistochemical evaluation of 15 Axon AT, Classen M, Cotton PB, et al. Pancreatography in chronic K-ras, p53, and HER-2/neu expression in hyperplastic, dysplastic, and pancreatitis: international definitions. Gut 1984;25:1107–12. carcinomatous lesions of the pancreas: evidence for multistep 16 Wynder EL, Mabuchi K, Maruchi N, et al. Epidemiology of cancer of the carcinogenesis. Hum Pathol 1999;30:123–9. pancreas. J Natl Cancer Inst 1973;50:645–67. 28 Ekbom A, Helmick C, Zack M, et al. Ulcerative colitis and colorectal 17 Lin RS, Kessler II. A multifactorial model for pancreatic cancer in man. cancer. A population-based study. N Engl J Med 1990;323:1228–33. Epidemiologic evidence. JAMA 1981;245:147–52. 29 Williamson WA, Ellis FH Jr, Gibb SP, et al. Barrett’s esophagus. 18 Gold EB, Gordis L, Diener MD, et al. Diet and other risk factors for Prevalence and incidence of adenocarcinoma. Arch Intern Med cancer of the pancreas. Cancer 1985;55:460–7. 1991;151:2212–16. 19 Mack TM, Yu MC, Hanisch R, et al. Pancreas cancer and smoking, 30 Lowenfels AB, Maisonneuve P, DiMagno P, et al. Hereditary beverage consumption, and past medical history. J Natl Cancer Inst pancreatitis and the risk of pancreatic cancer. International Hereditary 1986;76:49–60. Pancreatitis Study Group. J Natl Cancer Inst 1997;89:442–6. 20 Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. Aspects of 31 Gold EB, Cameron JL. Chronic pancreatitis and pancreatic cancer. N medical history and exocrine carcinoma of the pancreas: a Engl J Med 1993;328:1485–6. population-based case-control study in the Netherlands. Int J Cancer 32 Lowenfels AB. Pancreatitis and the risk of pancreatic cancer. N Engl J 199252:17–23. Med 1993;329:1502–3.

Reference linking to full text http://gut.bmj.com/ of more than 200 journals on September 28, 2021 by guest. Protected copyright.

Toll free links

You can access the FULL TEXT of articles cited in Gut online if the citation is to one of the more than 200 journals hosted by HighWire (http://highwire.stanford.edu) without a subscription to that journal. There are also direct links from references to the Medline abstract for other titles.